Documentation

Atlas.ComplexVariables.code.TaylorComplex

Taylor's Theorem for Complex Analytic Functions #

Formalization of Theorem 1 (Taylor's Theorem) from Lecture 10 of MIT 18.112 (Functions of a Complex Variable, Fall 2008).

Main results #

Strategy #

The remainder function is defined as the n-fold iterated divided slope (swap dslope a)^[n] f. The key ingredients are:

  1. dslope_expansion_identity — algebraic identity relating f to iterated dslope values.
  2. iterate_dslope_differentiableOn — iterated dslope preserves differentiability on open sets, using differentiableOn_dslope from the removable singularity theory.
  3. iterate_dslope_eq_iteratedDeriv_div_factorial — connects dslope coefficients to iteratedDeriv k f a / k! via formal power series machinery.
noncomputable def complexTaylorPoly (f : ) (a : ) (n : ) (z : ) :

The complex Taylor polynomial of degree n - 1 for f centered at a: ∑_{k < n} (f^(k)(a) / k!) (z - a)^k.

Instances For
    theorem dslope_expansion_identity (f : ) (a z : ) (n : ) :
    f z = kFinset.range n, (Function.swap dslope a)^[k] f a * (z - a) ^ k + (Function.swap dslope a)^[n] f z * (z - a) ^ n

    Algebraic identity: for any function f, point a, and evaluation point z, f(z) = ∑_{k<n} (swap dslope a)^[k] f(a) · (z-a)^k + (swap dslope a)^[n] f(z) · (z-a)^n. This is a purely algebraic consequence of the definition of dslope.

    theorem iterate_dslope_differentiableOn {f : } {Ω : Set } {a : } ( : IsOpen Ω) (ha : a Ω) (hf : DifferentiableOn f Ω) (n : ) :

    The n-fold iterated dslope preserves differentiability on open sets. Uses the complex removable singularity theorem (differentiableOn_dslope).

    The coefficient identity: if f has a formal power series expansion at a, then the value of the k-th iterated dslope at a equals f^(k)(a) / k!.

    theorem taylor_complex {f : } {Ω : Set } {a : } (n : ) ( : IsOpen Ω) (ha : a Ω) (hf : DifferentiableOn f Ω) :
    ∃ (fₙ : ), DifferentiableOn fₙ Ω zΩ, f z = complexTaylorPoly f a n z + fₙ z * (z - a) ^ n

    Taylor's Theorem (Lecture 10, Theorem 1, Part 1). If f is analytic (holomorphic) in a region Ω containing a, then for every n, $$f(z) = f(a) + \frac{f'(a)}{1!}(z-a) + \cdots + \frac{f^{(n-1)}(a)}{(n-1)!}(z-a)^{n-1}

    • f_n(z)(z-a)^n,$$ where fₙ is analytic (differentiable) in Ω.
    theorem circleIntegral_inv_pow_sub_mul_sub_eq_zero {a : } {R : } (hR : 0 < R) (z : ) (hz : z Metric.ball a R) (n : ) :
    (ζ : ) in C(a, R), ((ζ - a) ^ (n + 1) * (ζ - z))⁻¹ = 0

    Vanishing integral: ∮ 1/((ζ-a)^(n+1) * (ζ-z)) dζ = 0 for z ∈ ball(a, R) and n ≥ 0. Proved by induction on n using partial fractions and the residue computation ∮ (ζ-w)^m dζ = 0 for m ≠ -1.

    theorem taylor_integral_representation {f : } {Ω : Set } {a : } {R : } (n : ) ( : IsOpen Ω) (ha : a Ω) (hf : DifferentiableOn f Ω) (hR : 0 < R) (hball : Metric.closedBall a R Ω) (z : ) (hz : z Metric.ball a R) :
    (Function.swap dslope a)^[n] f z = (2 * Real.pi * Complex.I)⁻¹ * (ζ : ) in C(a, R), f ζ / ((ζ - a) ^ n * (ζ - z))

    Integral representation of the Taylor remainder (Lecture 10, Theorem 1, Part 2). If C is the boundary of a closed disk of radius R centered at a contained in Ω, then the remainder function fₙ satisfies $$f_n(z) = \frac{1}{2\pi i} \oint_C \frac{f(\zeta)\,d\zeta}{(\zeta - a)^n (\zeta - z)}$$ for z inside C.

    The proof uses Cauchy's integral formula and induction on n. The key auxiliary result is circleIntegral_inv_pow_sub_mul_sub_eq_zero, a vanishing integral proved by partial fractions.